sign in sign up
<<
Home , Social simulation, Systems theory

Social : Distinguishing Reflexive and Non-reflexive Modes

Christopher Goldspink Robert Kay Centre for Research in Social Head of Strategic Innovation, Department of Westpac Banking Corporation, University of Surrey. Guildford, GU1 7XH, UK Level 8, Westpac Place, 275 Kent Street, Sydney NSW 2000 [email protected] [email protected]

Abstract within the context of human social by focusing on Emergence has a long and controversial history. In this paper processes of ‘normative’ self-. The aim is to we briefly review the primary strands of the debate, paying contribute both to the conceptualization of emergence as attention to its use in the fields of of and well as to how social emergence may be meaningfully mind, and systems including the theory modeled. of complex systems. We argue that it is important to recognize We argue that the ambiguity, opaqueness and lack of why emergence in social systems is fundamentally different specification of the of emergence currently present a from other natural systems. The key characteristics of significant barrier to its application to the study of social reflexivity are discussed and a distinction between two systems. Furthermore we argue that social systems classes of emergence proposed. Non-reflexive emergence: represent a specific class of , distinct from other where the agents in the system under study are not self-aware, and Reflexive emergence: where the agents in the system under natural systems where emergence may be studied. This is study are self-aware and linguistically capable. We specify the due to the capability of human agents to distinguish ‘self’ generative processes we believe are associated with each of from ‘other’ and in doing so reflexively distinguish and these categories and argue for the adoption of this distinction interact with their , greatly increasing the in both theoretical and practical modeling of human social scope and of the emergent structures which are systems. possible. As a consequence we argue that both the form and mechanisms through which emergence occurs are not entirely analogous between natural and social systems. Introduction In this paper we review the historical and contemporary definitions of emergence, paying particular attention to its The concept of emergence has become widely used within use in the fields of and mind, social the social simulation community. The concept continues to be vaguely defined and to stand in for different science, general and complexity theory. This is followed by a discussion of the distinct propositions about social generative mechanisms. Within characteristics of social systems and the implications of the social simulation community, the concept has focused this difference for social simulation. The key primarily on upward causation (consistent with its usage characteristics of reflexivity will then be discussed and a within complex systems theory and associated research tentative framework will be proposed for two classes of programs such as those into artificial ) (Sawyer, 2003). Few attempts have been made to reconcile this use of the emergence, specifically: concept with its wider philosophical use and with the ∞ Non-reflexive emergence: where the agents in the parallel debates about the micro-macro link and the system under study are not self-aware, and relationship between structure and agency within the social . Relatively little attempt has been made to ∞ Reflexive emergence: where the agents in the system identify the defining characteristics of human social under study are self-aware and linguistically capable. systems and to critically re-examine the concept within this context. Similarly derivative such as We then specify the generative processes associated with downward causation and ‘immergence’ (Castelfranchi, each of these classes. 1998b) have only recently begun to be explored in the In proposing these two classes we do not preclude the simulation of human social systems. One current attempt distinction of other specific forms of emergence, but seek to to advance our understanding of upward and downward highlight the need for differing approaches to the study of causation in social systems is the EU funded project different system types, with a view to enhancing the Emergence in the Loop (EMIL). EMIL is concerned to explanatory power and applicability of the emergence explicate the mechanisms of emergence and immergence concept to various system classes.

Copyright © 2007, Association for the Advancement of (www.aaai.org). All rights reserved.

48 A Brief History of the Concept of Emergence 5. May be capable of top-down causation. 6. Are characterized by multiple realization or wild The notion of emergence has a long history, having been disjunction (Fodor, 1974) (alternative micro-states invoked in a number of disciplines with varying degrees of may generate the same macro states). to the theoretical and methodological development of associated fields. The concept remains A key concept is supervenience: a specification of the ambiguous and contentious, covering: ‘loose’ determinisms held to apply between levels such that ‘…an entity cannot change at a higher level without …a wide spectrum of ontological commitments. also changing at a lower level’ (Sawyer, 2001: 556). According to some the emergents are no more than , with no causal powers of their own; for Within this stream prominence of place is given to both downward and upward causation. Clayton and Davies others they are substances in their own right… (2006) specify downward causation as involving macro (Clayton, 2006: 14). structures placing constraint on lower level processes

hence ‘Emergent entities provide the context in which The first explicit use of the concept has been attributed to local, bottom up causation takes place and is made George Henry Lewes, in 1875 (Ablowitz, 1939). Following Lewes the concept rose to prominence primarily possible’ (Peterson, 2006: 697). This concept appears similar to that of ‘immergence’ within the social within the philosophy of science but more recently can be simulation literature and is worth exploring a little more seen to have been advanced within three distinct streams: fully as it is otherwise absent within the approach to philosophy, particularly philosophy of mind; systems emergence typical of complex systems inspired approaches theory, in particular complex systems; and social science (Sawyer, 2003, 2005). where it has largely been referred to under the heading of the micro-macro link and/or the problem of structure and Davies (2006) argues that the mechanism of downward causation can usefully be considered in terms of agency. Interestingly there has been relatively little cross boundaries. Novelty, he argues, may have its origin in a influence between these streams. While it is beyond the system being ‘open’. If novel order emerges it must do so scope of this paper to present a full comparison or to within the constraints of . He concludes: attempt a synthesis of the different streams, some brief comments are offered on the alternative perspectives and … top-down talk refers not to vitalistic augmentation contribution of each to the wider debate. of known forces, but rather to the system harnessing existing forces for its own ends. The problem is to The Contribution from Philosophy of science understand how this harnessing happens, not at the level of intermolecular interactions, but The philosophy of science and philosophy of mind stream overall – as a coherent project. It appears that once a is arguably the oldest – some date it back to system is sufficiently complex, then new top down (Peterson, 2006) but the debate is widely seen as having come to focus with the British Emergentists (Eronen, rules of causation emerge (Davies 2006: 48).

2004; Shrader, 2005; Stanford Encyclopaedia of For Davies then, top-down causation is associated with Philosophy, 2006). This school sought to deal with the self-organization. For Davies it is the ‘openness’ of some apparent qualitatively distinct associated with systems that ‘provides room’ for self-organizing process to different phenomena (physical, chemical, biological, arise, but he concludes, ‘openness to the environment mental) in the context of the debate between mechanism and vitalism: the former being committed to Laplacian merely explains why there may be room for top-down causation; it tells us nothing about how that causation causal determinism and hence and the latter works.’ The devil then, is in the detail of the mechanisms invoking ‘non-physical’ elements in order to explain the specific to particular processes in particular contexts and qualitative difference between organic and in-organic particular phenomenal domains. matter. This stream remains focused on explaining different properties of classes of natural phenomena and with the relationship between brains and minds (See Clayton & The contribution from Social Science Davies, 2006 for a recent summary of the positions). The micro-macro problem – the relationship between the Peterson (2006: 695) summarizes the widely agreed actions of and resulting social structures and characteristics of emergent phenomena within this stream the reciprocal constraint those structures place on as follows. Emergent entities: individual agency – has long standing in social science. 1. Are characterized by higher-order descriptions (i.e. The problem is central to many social developed form a ). throughout the 19th and 20th century. Examples include: 2. Obey higher order laws. Marxian dialectical materialism (Engels, 1934) built upon 3. Are characterized by unpredictable novelty. by, among others, Vygotsky (1962) and Lyont’ev (1978); 4. Are composed of lower level entities, but lower level the social constructionism of Berger and Luckmann entities are insufficient to fully account for emergent (1972); Gidden’s structuration theory (1984); and the entities (irreducibility). recent work of critical realists (Archer, 1998; Archer,

49 Bhaskar, Ciollier, Lawson, & Norrie, 1998; Bhaskar, ∞ A concern with boundaries and boundary conditions. 1997, 1998). These alternative theories are frequently founded on differing assumptions, extending from the More recently the development of complex systems theory essentially objectivist/rationalist theory of Coleman and its application to natural, social and cognitive (1994), through the critical theories of Habermas and then phenomena has provided additional concepts upon which to the radical constructivism of Luhmann (1990; 1995). much current debate about emergence draws. Many of these Fuchs & Hofkirchner (2005: 33) classify into four concepts and methods have become widely used within categories the ontological position of alternative approaches the multi-agent modeling community (Castelfranchi, to the micro-macro relationship. The majority of existing 1998a; Gilbert, 1995; Holland, 1998). social theory, they argue, fall into one or other of the first Within contemporary debate, and in contrast to the two categories: individualism and sociologism. Neither of position taken by the British Emergentists who argued that these ‘’ provides a theoretical foundation which irreducibility was the exception (Eronen, 2004), most real supports exploration, let alone advance understanding, of world systems are now argued to be non-linear (Kauffman, the interplay between agency and structure. The third 2000; Kauffman, 1993, 1996; Stewart, 1990). It is non- category, dualism, was the target of the original linearity which contributes to these system’s capacity for emergentists. Only those theories categorized as dialectical novelty and unpredictability in principle, through the therefore have relevance. Even here, it is reasonable to presence of deterministic Chaos (Lorenz, 2001; Williams, conclude that little practical advance has been achieved, as 1997) and/or equifinality. Equifinality as it is known most positions result in a straddling of bottom up and top- within systems theory, or the principle of ‘wild down arguments and/or suffer from excessively vague disjunction’ as it is known in philosophy, refers to a conceptualisation. What has been largely agreed, despite system where a single high level may be realized the very different theoretical and often inadequate handling by more than one set of micro-states which have no lawful of this problem, is that structure and agency come together relationship between them (Richardson, 2002b, 2002a; in activity or in bodyhood – the specific psycho-motor state Sawyer, 2001). As there is no a-priori basis by which the at the instant of enaction. Both Vygotsky and Giddens, for likely micro state can be determined, such systems are example, focus on action as the point of intersection irreducible and unpredictable in principle. between human agency and social structures. Summary The contribution from Systems Theory The concept of emergence has led to the establishment of a Systems language was clearly evident in the work of the number of general principles which describe the early emergentists and in a great deal of sociology and relationship between micro and macro phenomena, as well anthropology – notably that of and Gregory as some methods and techniques for identifying and Bateson. However, ‘systems’ as a focus of systematic exploring it. Specifically, we can conclude that there are research arguably took form with Bertalanffy’s attempt to systems which: establish a Theory in 1950 (Bertalanffy, 1950). As the science of ‘wholes’ systems theory stands in ∞ are inherently analytically reducible (to which the contrast to reductionisms concern with parts. In many concept of emergence does not apply); respects systems theory was put forward as a counter to ∞ are analytically reducible in principle but difficult to what was perceived as excessive reductionism dominating th reduce in practice and/or where an advance in scientific discourse during much of the 20 century. science/knowledge is needed for reduction to be While in the early stages of development of the theory, possible because the results were ‘unexpected’ systems tended to be modeled as ‘black boxes’ (effectively (Chalmers, 2006) (to which the concept of ‘weak’ masking the relationship between micro and macro emergence can be applied); elements), the application of the concept to social science, in particular through the development of social ∞ are not reducible in principle (to which the principle (Keeney, 1987) and soft systems approaches (Checkland, of ‘strong’ emergence is relevant). 1988) provided a theoretical lense as well as methods To which of these three classes do human social systems useful for describing the systemic behavior of social belong? We argue it is always the latter and our reason for systems. While the aspiration of GSM to establish a arguing this, and a discussion of the implications, are general science of systems is generally regarded to have taken up below. In the end, there is no substitute for failed (Jackson, 2000), systems approaches have attempting to specify the mechanism. In the following contributed valuable methods for the study of the interplay section we outline mechanisms fundamental to the between levels. The Systems view of emergence was emergence of social order and in so doing, attempt to founded on: clarify where and in what way these mechanisms rest on and differ from those which generate emergent ∞ ; the whole is greater than the sum of its parts. order in non-human natural systems. Throughout the ∞ A concern with both positive and negative. discussion, pointers are provided to where the mechanisms being outlined have, at least in part, been incorporated into

50 of artificial intelligence or artificial Non-reflexive social emergence . The pre-linguistic mode operates through the mechanism of structural coupling between agents. Structural coupling will arise between biological (autopoietic) agents which Towards a specification of the Generative Processes have sufficient cognitive range (behavioral repertoire) if they Particular to Human Social Systems are located in a common environment. Assuming that their phylogeny and ontogeny is such that they can co- Our ontological starting position is that physics constrains exist through the process of recurrent mutual perturbation, chemistry; chemistry constrains biology and biology, sociality – in other words we advocate a form of each will adjust its structure so as to accommodate the other – their structures will become mutually aligned or naturalism. However, at each of these ontological levels, structurally coupled. This process has been approximated novelty occurs due to the emergence of unique macro- in a simulation by Stoica-Kluver and Kluver (2006). configurations within the space of possibilities allowed by An observer may notice regularities in the resulting the lower level. There is downward causation through the patterns of interaction and these may be labeled as ‘norms’ way in which higher order patterns change boundary conditions at lower levels and constrain the situation for example. These ‘norms’ represent mutual accommodations, and an observer might attribute to those specific interactions that are possible. This suggests an accommodations some social ‘function’. The instantaneous co-penetration of levels (synchronic accommodations an agent makes to remain viable in one emergence) and also a sequential one (diachronic domain of interaction will need to be reconciled (within its emergence). body-hood) against accommodations being made The of the micro agents (their characteristics and action potentials), the heterogeneity and the structure of (simultaneously) as it also participates with different agents in other domain/s in which it is simultaneously their relationships will influence the range and type of participating – agency and structure converge and are both macro structures which can emerge. So far, though, there is instantiated at the point of enaction. The accommodations nothing here that is not true of any other natural system in made will be those that allow the agent to remain viable which there is the presence of non-linearity. What, then, and to maintain its organization (i.e. which ‘satisfice’ the are the fundamental characteristics of human agents, and what effect might the range and type of action potentials constraints and allow conservation of identity) based on its unique ontogeny (structure resulting from its history of have on emergent social structures? interactions in a variety of domains). Human agents are cognitive. Within the biological Here the emergent structure can be seen to be ‘in’ (i.e. sciences human cognition is recognized as involving internalized within its own cognitive structure) each agent activity, where external perturbation on to the extent that each has had to make structural sensory surfaces results in a set of behavioral dispositions and responses to the environment. Like many animals, adjustments to operate in the shared domain. The structural adjustment each needs to make in order to Humans form social systems by coordinating their persist will, however, be unique – in other words the behavior through reciprocal action. Unlike many other structural accommodations each has made in order to animals, Humans are capable of coordinating their contribute to the patterns we are calling ‘norms’ for coordination of action by way of language. The biologists example, will not be the same. The structure, then, can and developed a comprehensive theory of this process – the theory of also be regarded as ‘in’ the network, as it is the intersection of these disparate agent structures which gives (1980) and subsequently explored its it its particular form at a particular time. As any agent implications for human social systems (Maturana, 1988a; could leave the domain and have minimal effect on the Maturana, 1988b; Varela, 1981; Varela, 1987; Varela, resulting , each agent’s ‘contribution’ will be Thompson, & Rosch, 1992). Elsewhere, we have argued relatively small. The pattern that is labeled as ‘norm’ can that the theory of autopoiesis and its associated theory of enactive cognition is consistent with ideas be thought about as like a hologram. The whole is in every part (agent) such that removal of parts (agents) and that when combined with complexity, provides one reduces the resolution (coherence) but does not constitute possible pathway for understanding the substantive loss of overall pattern. However, the loss of too many mechanisms of sociality (Goldspink & Kay, 2003b; components may reduce the coupling to the point that the Goldspink & Kay, 2003a). As a part of this work, we existing pattern de-coheres and transforms into something argued that social systems represent a distinct class of complex systems and that this difference was significant in different. Each agent contributes to the , so it is conceivable that the pattern will only be realized terms of the range and type of emergent structures which it with some critical minimal number of agents present which implied. From this, two systemic mechanisms suggest have had a sufficient mutual history to have aligned their themselves as the generative source of emergent patterns structures (become socialized within that context). As which we commonly refer to as social structures – the first agents leave, the coherence may degrade until, beyond of these is pre-linguistic (and by definition non-reflexive) and the second involves language and hence is reflexive some critical point, it may de-cohere or take up an alternative ‘shape’ due to the influence of external (Gardenfors, 2006).

51 perturbations (acting through the remaining agents as dimensionality of the and of the resulting points of intersection of domains), or to the entry of new higher order structures it can generate and support. This is agents with different ontogeny. because language makes possible the emergence of domains Note that this emergence is consistent with that which we of interaction which can themselves become the target for would observe between any complex natural systems – the further linguistic distinction and hence new domains. In emergent pattern is the product of local interactions only. other words, language allows the agent to make This is relatively easy to model. While feedback from distinctions on prior distinctions (to language about its macro level to the micro is possible, it is only by means of prior language or to build further abstractions on prior effects of on the environment and then the abstractions). This supports the possibility of infinite environment on the individual – i.e. downward causation. recursion and infinite branching (there are no doubt In natural systems, the local level interactions between biological constraints on this in humans). Furthermore, a agents are constrained by the existing structures of the capacity to distinguish (label or categorize) processes agents and the state of their environment. With biological supports reification and this simplifies the cognitive agents the system is open in that any emergent structure is handling of processual phenomena and allows the resulting possible as long as it remains consistent with the reifications to be treated by the agent in the same manner biological viability of the agents as living (autopoietic) as material objects. entities. This biological constraint includes limits to These capabilities greatly expand the structural flexibility environmental conditions conducive to life (i.e. not too hot of the agents: they can now invent shared epistemic or too cold, the need for , limitations to sensory worlds. The phase space of agent cognition is now based channels, channel bandwidths and affective/psychomotor primarily on constraints of ontogeny rather than phylogeny response capabilities etc). These are primarily a product of and is hence under the influence of the agent/s. phylogeny (the evolutionary history of the organism at the Language makes possible a further major qualitative level of the species) rather than ontogeny (the history of difference in natural and human social emergence. Humans development at the level of the individual), and are (and possibly some other primates, cetaceans and therefore slow to change and not under the control of the elephants)1 have developed sufficient cognitive capacity to emergent . As a consequence the basic become self-aware and as such exhibit reflexive behavior. dimensionality of the phase space of the social system does This occurs when the agent is capable of distinguishing not change over the time frame of interest for understanding ‘self’ and ‘other’ i.e. the agent can entertain the notion of social systems. The dimensionality of the phase space is ‘I’ as a concept and treat that concept as an object. The determined by the dimensions of variability possible by advent of this capacity for reflexive identity also supposes individuals – i.e. the plasticity of their nervous systems the existence of a range of conceptual operators that act on and by higher order dimensions which emerge from their identity – identity construction and maintenance becomes a interaction. part of the agent’s world creation. This gives rise to what Gilbert has called second order emergence: Reflexive Social Emergence second order emergence occurs when the agents Our sensory surfaces are designed to detect difference in recognize emergent phenomena, such as societies, some dimension of the world and our cognitive apparatus clubs, formal , , localities is thus geared to make distinctions. Once our cognitive and so on where the fact that you are a member or a complexity exceeds a critical threshold (Gardenfors, 2006) non-member, changes the rules of interaction between these distinctions can be represented in language. Maturana you and other agents. (Gilbert, 2002: 6). and Varela (1980) describe language as involving the co- ordination of the co-ordination of actions – i.e. language In other words, agents can now notice the patterns that provides a meta process by which agents orientate arise as they interact with others and distinguish those themselves within a world. Structural coupling can arise patterns in language. For example, with the EMIL project purely through behavioral coordination of action (as which is concerned to understand the emergence and self- discussed above), but it can also take place in and through regulatory role of social norms, a reflexive agent can notice linguistic exchange – the mutual co-ordination of co- an emergent pattern of social behavior and explicitly denote ordination of behaviors. This gives rise to a consensual it as a ‘norm’. While this denotation may be idiosyncratic linguistic domain characterized by a more or less shared (i.e. based on the necessarily limited of the lexicon. This process has been simulated using both individual agent), the agent can nonetheless act on the shared referents and simple structural coupling in the basis of this denotation. Once distinguished and reified absence of objective referents (Gong, Ke, Minett, & Wang, within a domain, agents can decide (on the basis of rational 2004; Hutchins & Hazlehurst, 1995; Steels, 1997, 1998;

Steels, 2005; Steels & Kaplan, 1998; Steels & Kaplan, 1 It is important to note that we can infer the existence of threshold 1999), as has the emergence of a rudimentary grammar effects here but cannot precisely specify the critical points of (Howell & Becker, n.d; Vogt, n.d). complexity at which self- and language becomes possible. The advent of language radically increases the behavioral The ability for language is of course evident in species other than plasticity of agents and has significant implications for the humans, but the degree to which their linguistic plasticity involves or enables reflexivity in the system is a subject for further research.

52 as well as value based or emotional criteria) how to Alex Ryan (2006). In most social theory, positing the respond – they can choose to ignore the norm or to behave observer as a necessary part of the system removes any in ways they believe will limit the reoccurrence of the ontological privilege and threatens either infinite recursion behaviors that are outside the agreed/shared patterns of the or paradox. Based on the position advocated here, a degree group. Once a pattern has been distinguished in language it of both may well be fundamental to the type of system can make the transition to a rule – a formally stated, being described (Hofstadter, 2007). linguistically explicit requirement with stated conditionals and possible resources to maintain it. This suggests, for Implications for emergence example, that an agent can form hypotheses about the Complex systems demonstrate a capacity to give rise to relationship between a macro structural aspect of the social system in which it is a participant and then act on that complex macro patterns as a result of local interactions between agents in highly connected webs. This local hypothesis, potentially changing the structure which it interaction can often be characterized as involving some participates in generating. This gives rise to a feedback signaling between agents. As we have seen above, in path between macro and micro phenomena that is not human social systems, this signaling behavior takes on a present in any other natural phenomena. qualitatively different form. This has three key implications The recognition that agents possessing this cognitive complexity form the components of a social system sets up for our understanding of emergence that to date have largely been ignored by the literature. a distinct class of emergence. This is on the basis that reflexive agents will display qualitatively different 1. Social systems will display an increased range of behaviors from non-reflexive through the ability to modify emergent possibilities: the reflexive nature of social their own sets of behavioral change triggers. For agents systems implies that a greater range of emergent which have linguistic capability, the two processes structures should be expected and they will be subject (linguistic and non-linguistic) intertwine or even become to more rapid change. one and would not be able to be empirically disentangled. Their respective influences will only be able to be 2. Dimensions of phase space are non-constant: As the examined through simulations or by comparing agents agents in the social system define and redefine the with different (phylogenetic) capabilities (i.e. different phase space as a function of their reflexive distinctions species). they will create and change the dimensions of that phase space, in order to support their own viability in that space. The Role of the Observer Another significant implication of the relationships 3. Phase space comes under control of the system and is described above is the observer dependant nature of dynamic: The dimensionality of the phase space emergence in social systems. In human social systems associated with ontogenetic parameters is derived every agent is an observer and it is the process of through the self-distinguishing characteristics of the observation and the associated distinction-making which is agents and can be influenced by their situated the reflexive engine of emergence. In natural systems, the behavior. agents of the system are unable to observe and distinguish linguistically or to distinguish external structures as separate from themselves hence the process of observation Conclusion has no impact on the dynamics of the system or the way in In this paper we have argued that the notion of emergence which emergence takes place. To some extent we can see as it is currently discussed in the literature fails to consider an acknowledgement of this effect in methodological that alternative generative mechanisms, and hence forms of discussions within ethnography, action research (Carr & emergence, apply to different classes of phenomena. Kemmis, 1986) and grounded theory (Corbin & Strauss, Complex systems, and artificial societies 1990). In each of these methodologies the impact of the currently model bottom up emergence and systems where researcher on the social system under study is top down influence operates only indirectly by downward acknowledged and seen as part of the process. The view propagation of constraint, not by more direct feedback. We being proposed here is that any agent that becomes a part have argued that this is not adequate for an understanding of the system being observed has the potential to influence of human social systems. We have set out two that system. An agent can become a part of the system mechanisms present in human social systems – non- simply by being itself observed or conceived as observing reflexive and reflexive; suggested a suitable theoretical by those who constitute the system. In other words, the frame from which they may be considered – that of effect of the entry of a new observing agent is to change the autopoiesis; and examined some of the implications these system boundary so as to include that agent. The boundary alternatives may have for the behavior of systems which is itself an entity of ambiguous status – it is an epistemic support them. distinction albeit one based on potentially ontological We have argued that autopoietic human agents display a markers. The most elegant handling of this concept we qualitatively different range of behavioral possibilities have encountered in the context of emergence is that of brought about by their distinctive (biological) properties,

53 primarily an advanced nervous system which supports the Clayton, P. & Davies, P. 2006. The Re-Emergence of development of language and a capacity for reflexive self- Emergence: The Emergentist Hypothesis from Science to awareness. These broadened behavioral capabilities, in Religion. Oxford: Oxford University Press. particular language, imply that the process by which Coleman, J. S. 1994. Foundations of Social Theory. Cambridge: emergence occurs in social systems is fundamentally Belknap. Corbin, J. M. & Strauss, A. 1990. Grounded theory research: different to that which is observed in other systems Procedures, canons, and evaluative criteria. Qualitative populated by non-linguistic agents. Sociology, 13(1): 18. We suggest that this framework may support a more Davies, P. 2006. The Physics of Downward Causation. In P. focused research program that doesn’t unnecessarily mix Clayton & P. Davies (Eds.), The Re-Emergence of different classes of system under the same Emergence: The Emergentist Hypothesis from Science to conceptualization. At the same time we acknowledge the Religion. Oxford: Oxford University Press. possibility for yet further classes of emergence to be Engels, F. 1934. Dialectics of Nature. Moscow: Progress identified due to a current lack of understanding within the Publishers. literature on the threshold points between system classes as Eronen, M. 2004. Emergence in the Philosophy of Mind. they relate to emergence. We suspect, for example, that the University of Helsinki, Helsinki. Fodor, J. A. 1974. Special; Sciences or The Disunity of Science locutionary, illocutionary and perlocutionary aspects of as a Working Hypothesis. Synthese, 28: 18. language identified by linguists may imply different Fuchs, C. & Hofkirchner, W. 2005. The Dialectic of Bottom-up mechanisms. and Top-down Emergence in Social Systems. tripleC 1(1): 22. Acknowledgements: One of the authors is involved with the Gardenfors, P. 2006. How Homo became Sapiens: On the project EMIL, “Emergence In the Loop: simulating the two of Thinking. Oxford: Oxford University Press. way dynamics of norm innovation”, a three-year EU funded Giddens, A. 1984. The Constitution of : Outline of the project (Sixth Framework Programme - Society and theory of structuration. Berkeley: University of California - Citizens and Governance in the Knowledge Press. Based Society). The views expressed are the authors and do Gilbert, N. 1995. Artificial Societies. In N. Gilbert & R. Conte not necessarily reflect those of the Partners to this project. (Eds.), Emergence in Social Simulation. London: UCL Press. Gilbert, N. 2002. Varieties of Emergence. Paper presented at References the Social Agents: , Exchange, and Evolution Conference Social agents: ecology, exchange, and evolution, Ablowitz, R. 1939. The Theory of Emergence. Philosophy of Chicago Science, 6(1): 16. Goldspink, C. & Kay, R. 2003a. Organizations as Self Archer, M. 1998. Realism in the Social Sciences. In M. Archer Organizing and Sustaining Systems: A Complex and & R. Bhaskar & A. Collier & T. Lawson & A. Norrie (Eds.), Autopoietic Systems Perspective. International Journal Critical Realism: Essential Readings. London: Routledge. General Systems, 32(5): 459-474. Archer, M., Bhaskar, R., Ciollier, A., Lawson, T., & Norrie, A. Goldspink, C. & Kay, R. 2003b. Bridging the Micro-Macro 1998. Critical Realism: Essential Readings. London: Divide: a new basis for social science. Human Relations, 57 Routledge. (5): 597-618 Berger, P. L. & Luckman, T. 1972. The Social Construction of Gong, T., Ke, J., Minett, J. W., & Wang, W. S. 2004. A Reality: Penguin. Computational Framework to Simulate the co-evolution of Bertalanffy, L. v. 1950. An Outline of General Systems Theory. language and . British Journal for the Philosophy of Science, 1(2). Hofstadter, D. R. 2007. I am a Strange Loop: Basic Books. Bhaskar, R. 1997. A Realist Theory of Science. London: Verso. Holland, J. H. 1998. Emergence: from chaos to order. Ma.: Bhaskar, R. 1998. The Possibility of Naturalism. London: Addison Wesley. Routledge. Howell, S. R. & Becker, S. n.d. Modeling Language Carr, W. & Kemmis, S. 1986. Becoming Critical: Knowing Acquisition: Grammar from the Lexicon? Through Action Research: Deakin University. Hutchins, E. & Hazlehurst, B. 1995. How to invent a lexicon: Castelfranchi, C. 1998a. Simulating with Cognitive Agents: the development of shared symbols. In N. Gilbert & R. Conte The Importance of Cognitive Emergence. In J. S. Sichman & (Eds.), Artificial Societies. London: UCL Press. R. Conte & N. Gilbert (Eds.), Lecture Notes in Artificial Jackson, M. C. 2000. Systems Approaches to . Intelligence. Berlin: Springer Verlag. London: Kluwer Academic. Castelfranchi, C. 1998b. Through The Minds of the Agents. Kauffman, S. 2000. Investigations. New York: Oxford. Journal of Artificial Societies and Social Simulation, 1(1). Kauffman, S. A. 1987. The Evolution of Economic Webs. In P. Chalmers, D. J. 2006. Strong and Weak Emergence. Canberra: W. Anderson & et al. (Eds.), The Economy as an Evolving Research School of Social Sciences, Australian National Complex System: Addison-Wesley. University, Canberra, Australia. Kauffman, S. A. 1993. The Origins of Order: Self Organization Checkland, P. 1988. Systems Thinking Systems Practice. G.B.: and Selection in Evolution: Oxford University Press. John Wiley. Kauffman, S. A. 1996. At home in the Universe: The Search for Clayton, P. 2006. Conceptual Foundations of Emergence Laws of Complexity. London: Penguin. Theory. In P. Clayton & P. Davies (Eds.), The re-Emergence Keeney, B. P. 1987. Aesthetics of Change, N.Y.: Guilford of Emergence: The Emergentist Hypothesis from Science to Press. Religion. Oxford: Oxford University Press.

54 Kennedy, J. & Eberhart, R. C. 2001. Swarm Intelligence (1 ed.). Stioica-Kluver, C. & Kluver, J. 2006. Interacting Neural London: Academic Press. Networks and the Emergence of Social Structure. Leont'ev, A. N. 1978. Activity, Consciousness and Complexity, 12(3): 11. Personality. Engelwood Cliffs: Prentice Hall. Varela, F. 1981. Describing the logic of the living. The Lorenz, E. N. 2001. The Essence of Chaos (4 ed.). Seattle: adequacy and limitations of the idea of Autopoiesis. In M. University of Washington Press. Zeleny (Ed.), Autopoiesis: A theory of the living Luhmann, N. 1990. Essays on Self Reference. New York: organisation. New York: Elsevier-North Holland. Columbia University Press. Varela, F., Thompson, E., & Rosch, E. 1992. The Embodied Luhmann, N. 1995. Social Systems. Stanford: Stanford Mind. Cambridge: MIT Press. University Press. Varela, F., ', In (ed.) , ,. 1987. Laying Down a Path in Walking. Maturana, H. & Varela, F. 1980. Autopoiesis and Cognition: In T. W.I. (Ed.), GAIA - A Way of Knowing. San Francisco: The Realization of the Living. Boston: D. Reidel. Indisfarne Press. Maturana, H. 1988a. The of Observing: The Vogt, P. n.d. Group Size Effects on the Emergence of biological Foundations of Self-Consciousness and the Compositional Structures in Language. Tilburg, Physical Domain of Existance. Paper presented at the Netherlands: Tilburg University. American Society for Cybernetics Conference, Felton CA. Vygotsky, L. S. 1962. Thought and Language. Cambridge, Maturana, H. R. 1988b. Reality: The Search for Objectivity of Mass: MIT Press. the Quest for Compelling Argument. Irish Journal of Williams, G. P. 1997. Tamed. Washington D.C: , 9(1). Joseph Henry Press. Peterson, G. R. 2006. Species of Emergence. Zygon, 41(3): 22. Prigogine, I. 1997. The End of Certainty: Time, Chaos and the New Laws of Nature. New York: The Free Press. Richardson, K. A. 2002a. On the Limits of Bottom Up : Towards a Non-linear Modeling Culture. Paper presented at the 36th Hawaii International Conference on , Hawaii. Richardson, K. A. 2002b. Methodological Implications of a Complex Systems Approach to Sociality: Some further remarks. Journal of Artificial Societies and Social Simulation, 5(2). Ryan, A. 2006. Emergence is Coupled to Scope not Level. Sawyer, K. R. 2001. Emergence in Sociology: Contemporary Philosophy of Mind and Some Implications for Sociology Theory. American Journal of Sociology, 107(3): 34. Sawyer, K. R. 2003. Artificial Societies: Multi-agent Systems and the Micro-macro Link in Sociological Theory. Sociological Methods & Research, 31: 38. Sawyer, K. R. 2005. Social Emergence: Societies as Complex Systems. Cambridge, UK: Cambridge University Press. Shrader, W. E. 2005. The Metapysics of Ontological Emergence. University of Notre Dame. Stanford Encyclopedia of Philosophy. 2006. Emergent Properties, Stanford Encyclopedia of Philosophy. Steels, L. 1997. Constructing and Sharing Perceptual Distinctions. Paper presented at the European Conference on , Berlin. Steels, L. 1998. Structural coupling of cognitive memories through adaptive language games. Paper presented at the fifth international conference on simulation of adaptive behavior on From animals to animats 5, Univ. of Zurich, Zurich, Switzerland. Steels, L. & Kaplan, F. 1998. Stochasticity as a Source of Innovation in Kanguage Games. In C. Adami & R. K. Belew & H. Kitano & C. Taylor (Eds.), Artificial Life VI. Cambridge, MA: MIT Press. Steels, L. & Kaplan, F. 1999. Bootstrapping grounded word semantics. In T. Briscoe (Ed.), Linguistic evolution through language acquisition: formal and computational models,. Cambridge, UK: Cambridge University Press. Steels, L. 2005. The emergence and evolution of linguistic structure: from lexical to grammatical systems. Connection Science, 17(3 & 4): 17. Stewart, I. 1990. Does God Play Dice - The New Mathematics of Chaos: Penguin.

55